Explore CPU speed benchmarks of small language models
Explore model performance on the BananaMind benchmark
Chat with KeyLM here to test it out!
A 5M hybrid-looped LLM with effort levels